ACLS Algorithms Explained: What You Are Actually Deciding at Each Branch
Every code cart in the hospital has the same laminated ACLS poster taped to the top, and almost nobody is looking at it when it matters. That is fine. The poster is a reference, not a script. If you understand what each branch of the algorithm is actually asking you to decide, you do not need the boxes and arrows. You need two questions, a watch, and a team that talks to each other.
This post walks through the arrest algorithm, the bradycardia and tachycardia branches, and the part nobody prints on a poster: what the person running the code should be saying out loud. It is the conceptual companion to my full guide on running a code, which covers the operational side: who stands where, how to take over a code that is falling apart, and what to document afterward. Read that one too. This page is about the decisions. That page is about the room.
The two questions underneath every arrest
Strip away everything else and cardiac arrest management reduces to two questions you ask over and over until the patient gets a pulse back or you call it.
Question one: is the rhythm shockable? Ventricular fibrillation and pulseless ventricular tachycardia get electricity. Electricity is the single most effective therapy in the entire algorithm, and it only works on those two rhythms. Every rhythm check exists to answer this one question. Pulseless electrical activity and asystole do not get shocked, no matter how much everyone in the room wants to do something dramatic.
Question two: is there a reversible cause I have not fixed yet? Compressions and epinephrine buy time. They do not fix anything. If the arrest has a cause you can reverse, the patient survives when you find it and dies when you do not. This question matters most in PEA and asystole, where it is the only question, but it applies to every arrest.
The scaffolding that holds those two questions together is stable and worth memorizing cold: high quality compressions in 2 minute cycles, a rhythm check at the end of each cycle, and epinephrine 1 mg IV every 3 to 5 minutes for as long as the patient remains pulseless. Where the finer details sit, exact drug sequence, energy settings, airway timing, confirm against the current AHA algorithms and your hospital protocol. The two questions do not change between guideline cycles. The trivia does.
The VF and VT loop is a rhythm of decisions, not a poster
The shockable branch looks busy on paper. At the bedside it is a loop with a beat to it: shock, compress for 2 minutes, check the rhythm, decide again. In the standard sequence, epinephrine enters after the loop is established and an antiarrhythmic such as amiodarone comes in for VF or VT that keeps surviving shocks. Confirm the exact order and doses against the current AHA algorithms. What you should carry in your head is not the sequence trivia. It is the three decisions that repeat every cycle.
First: charge before you pause. The defibrillator should be charging during the last seconds of compressions so the rhythm check, the shock decision, and the shock itself happen inside one short pause. Hands-off time is the enemy. Every second off the chest drops coronary perfusion pressure, and it takes a long stretch of good compressions to build it back.
Second: call the rhythm out loud and commit. Shockable or not shockable. Say it. If it is VF, shock it and get back on the chest immediately. Do not stare at the monitor looking for a pulse in a rhythm that cannot produce one.
Third: ask what has changed. If the patient is still in VF after several shocks, the loop is telling you something. Refractory VF has causes: ischemia, potassium, magnesium, toxins. The loop keeps you organized, but it is not an excuse to stop thinking.
PEA and asystole: hunt the H's and T's, do not recite them
The nonshockable branch is humbling because the algorithm gives you almost nothing: compressions, epinephrine every 3 to 5 minutes, and a list. There is no shock to deliver. The entire game is finding the reversible cause, and this is where I watch teams fail. Someone at the foot of the bed recites hypoxia, hypovolemia, hydrogen ion, hypokalemia, hyperkalemia, hypothermia, tension pneumothorax, tamponade, toxins, thrombosis coronary, thrombosis pulmonary. Everyone nods. Nobody does anything. Reciting the list is not hunting.
Hunting means converting each item into a concrete check or an empiric action, assigned to a person, right now.
- Hypoxia: is the tube in the trachea, is the bag hooked to oxygen, is the chest rising. Confirm it. Do not assume it.
- Hypovolemia: squeeze in fluid wide open and ask the question out loud: is this patient bleeding? Recent surgery, GI bleed, anticoagulation.
- Hyperkalemia: this is the one that kills dialysis patients. If the patient has renal failure and arrested, give calcium and treat for hyperkalemia empirically. Do not wait for a lab value that will come back after the code is over.
- Tension pneumothorax: recent central line, recent intubation, chest trauma, absent breath sounds on one side. Decompress it.
- Tamponade: put the ultrasound on the heart if you have it and someone who can use it.
- Thrombosis: massive PE and coronary occlusion are the two big ones. The history you gathered before the code, or the story the nurse can give you in one sentence, is what points here.
Notice what makes this work: information. The team member who read the chart, the nurse who knows the patient missed dialysis Tuesday, the intern who remembers the potassium from this morning. This is also why so many arrests are preventable. Most inpatient codes announce themselves hours in advance as the kind of deterioration a rapid response call exists to catch. The best code is the one you never have to run.
Bradycardia and tachycardia: is the patient sick because of the rhythm?
The perfusing-rhythm branches confuse people because they look like rhythm algorithms. They are not. They are perfusion algorithms. The first decision in both is identical: is the patient unstable because of this rhythm? Hypotension, altered mental status, ischemic chest pain, acute heart failure. If yes, you treat the rhythm with electricity or pacing now. If no, you have time to think, and you should use it.
Bradycardia: a heart rate of 38 in a comfortable, warm, mentating patient is a finding. The same rate in a gray, confused, hypotensive patient is an emergency. For the unstable patient, atropine is the standard first move, with pacing and chronotropic infusions behind it. For the stable patient, your job is the cause: beta blockers, calcium channel blockers, hyperkalemia, ischemia, and conduction disease. This exact split shows up constantly on night calls, where the question is never what the atropine dose is. It is whether you need it at all.
Tachycardia: unstable gets synchronized cardioversion. Stable buys you time to answer two sorting questions: wide or narrow, regular or irregular. Narrow and regular points one direction, irregular another, and wide complex tachycardia gets treated as VT until proven otherwise. The classic mistake is giving a wide complex rhythm a drug meant for a narrow one because someone talked themselves into aberrancy. When you are not sure, assume the dangerous thing.
What the person running the code says out loud
The code leader's job is not to push drugs or do compressions. It is to run the loop out loud so thirty seconds of chaos never becomes five minutes of it. In practice that sounds like this:
"Rhythm check in 30 seconds. Charge the defibrillator." Then: "Pause. VF. Shocking. Resume compressions." Then: "That was our third shock, second epi at 14:32. Next rhythm check at 14:34."
Three things about that script. It announces the plan before the pause, so the pause is short. It keeps a running count of shocks and epinephrine doses, out loud, so the recorder and the leader agree. And once per loop in a nonshockable arrest, it opens the floor: "PEA. Somebody tell me an H or a T we have not addressed." That single sentence turns a room full of spectators into a search party.
Closed loop communication saves more lives than drug trivia
Here is the uncomfortable truth about ACLS: the drugs are the least important part. Epinephrine timing matters far less than compression quality, pause length, and whether orders actually get executed. A code where "give epi" floats into the air and three people assume someone else drew it up is a code where the epi never gets given.
Closed loop communication fixes this and it costs nothing. The order has a name attached: "Maria, give 1 mg epinephrine IV now." The receiver repeats it back: "1 mg epi going in." The receiver reports completion: "Epi is in." Three sentences. Every order, every time. Residents spend hours memorizing second line antiarrhythmics and thirty seconds practicing this, which is exactly backward. Nobody dies because the leader forgot a third line drug. People die because the first line drug was ordered into the air and never given.
Where to go from here
If you are studying this for an exam rather than the bedside, the branch points above are precisely what gets tested: shockable versus not, stable versus unstable, wide versus narrow. Our emergency medicine board review drills these decisions as question banks rather than posters, and the full board review library covers the rest of the exam the same way. If you want all of it plus the practice exams and the community, that is what membership is for. Learn the decisions, not the poster. The poster will change. The decisions will not.
Where HistoryandPhysical.net fits in
Understanding the branch points is the study version. The bedside version is being the person who knows what happens next.
- A free running-a-code guide. Running a code covers the role you are actually given, and the rapid response guide covers the call that comes before it.
- Emergency medicine question banks. An EM shelf bank and an EM board review bank for the exam side, with a full explanation on every answer.
- Written in 2026. Algorithms change. Every question in the catalog was written this year rather than inherited from a library that has been lightly edited for two decades.
- No ads, anywhere. Nothing on this site is trying to sell you something in the middle of a code checklist.
The guides are free forever. Membership opens the banks at $10 a month or $59 a year at the founding rate, with cancellation in two clicks.
Keep going with Alo Academy
Writing a good note is one part of the job. Passing the exams that get you there is another. Alo Academy covers those with question banks kept current with the exams, explanations that show the reasoning rather than the answer, and AI analysis that finds the specific mistakes costing you points, plus a dashboard that turns it into what to study next.